Call center outbound call completing rate improving method, system and device and storage medium

Abstract

The invention discloses a call center outbound call completing rate improving method, system and device and a storage medium, and relates to the technical field of call center management, and the method comprises the steps: collecting historical outbound call data of a call center; analyzing historical outbound data by using a machine learning algorithm, and establishing an outbound connection prediction model; based on an outbound call connection prediction model, calculating an optimal outbound call time window and an expected connection probability for each number to be called; formulating an outbound strategy according to the optimal outbound time window and the expected connection probability, and generating an outbound task queue according to the outbound strategy; executing an outbound task according to the outbound task queue, and generating an outbound task record; and updating the outbound call connection prediction model and the outbound call strategy by using the outbound call task record, dynamically optimizing the outbound call process, and completing the improvement of the outbound call connection rate of the call center. According to the method, the historical outbound data are analyzed through the machine learning algorithm, the outbound call connection prediction model is established, the prediction accuracy of the outbound call connection rate is improved, and the resource utilization efficiency is also greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of call center management, and in particular to a method, system, device and storage medium for improving the outbound call connection rate of a call center. Background Art

[0002] As an important channel for enterprises to communicate with customers, the operational efficiency of call centers directly affects the quality of customer service and the operating costs of enterprises. In the daily operations of call centers, outbound business accounts for a considerable proportion, and the outbound call connection rate is one of the key indicators to measure outbound call efficiency. However, traditional outbound methods often face problems such as low connection rate and low resource utilization, which affects the overall operational effect of call centers.

[0003] At present, most call centers still use fixed-time batch outbound calls or simple random dialing strategies. These methods fail to fully consider the individual characteristics and environmental factors of customers, resulting in a large number of invalid outbound calls and frequent interruptions to customers. At the same time, the lack of accurate prediction models and dynamic adjustment mechanisms makes it difficult for call centers to cope with the ever-changing outbound call environment and achieve optimal resource allocation. Summary of the invention

[0004] In view of the problems existing in the existing call center outbound calling method, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is how to improve the outbound call connection rate of the call center, optimize resource allocation, improve operational efficiency, and improve customer experience through intelligent algorithms and dynamic strategies.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In the first aspect, an embodiment of the present invention provides a method for improving the outbound call connection rate of a call center, which includes collecting historical outbound call data of the call center, the historical outbound call data including outbound call time, call duration, connection status, and user feedback information; using a machine learning algorithm to analyze the historical outbound call data and establish an outbound call connection prediction model; based on the outbound call connection prediction model, calculating the optimal outbound call time window and the expected connection probability for each number to be called; formulating an outbound call strategy based on the optimal outbound call time window and the expected connection probability, and generating an outbound call task queue based on the outbound call strategy; executing outbound call tasks according to the outbound call task queue, and generating an outbound call task record; using the outbound call task record to update the outbound call connection prediction model and the outbound call strategy, dynamically optimizing the outbound call process, and completing the improvement of the outbound call connection rate of the call center.

[0008] As a preferred solution of the method for improving the outbound call connection rate of the call center described in the present invention, the establishment of an outbound call connection prediction model includes the following steps: preprocessing historical outbound call data; extracting multidimensional features from the preprocessed historical outbound call data, the multidimensional features including time features, user features and environmental features; constructing training samples, taking the multidimensional features as input and the connection status as output labels; selecting multiple machine learning algorithms, and training each machine learning algorithm using the training samples to obtain multiple candidate models; using the verification data set to calculate the connection probability prediction accuracy of each candidate model, and selecting the candidate model with the highest connection probability prediction accuracy as the outbound call connection prediction model; and performing an explainability analysis on the outbound call connection prediction model.

[0009] As a preferred solution of the method for improving the outbound call connection rate of the call center described in the present invention, the following steps are included in the calculation of the optimal outbound call time window and the expected connection probability for each number to be called: collecting multi-dimensional information of the number to be called, the multi-dimensional information including the user's historical call records, user attributes and the current time; extracting features from the multi-dimensional information of the number to be called, generating a feature vector consistent with the input format of the outbound call connection prediction model; inputting the feature vector into the outbound call connection prediction model to obtain the expected connection probability at different time periods of the day; calculating the average expected connection probability of all time periods of the day, marking the time periods with a higher expected connection probability as candidate outbound call time windows; The candidate outbound call time windows are clustered and analyzed to identify and merge windows that are continuous in time to form the best outbound call time window. The maximum window duration threshold Tmax is set. If the duration of the best outbound call time window exceeds Tmax, the best outbound call time window is divided into N sub-windows of equal length, and the duration of each sub-window is less than or equal to Tmax. If there are multiple sub-windows, all of these sub-windows are regarded as the best outbound call time windows and are numbered in the original time sequence. The average expected connection probability within the best outbound call time window is calculated as the expected connection probability of the number to be called. The best outbound call time window, the expected connection probability and the number to be called are associated and stored.

[0010] As a preferred solution of the method for improving the outbound call connection rate of the call center described in the present invention, the formulation of the outbound call strategy includes the following steps: setting a first expected connection probability threshold and a second expected connection probability threshold; dividing the to-be-called numbers into high priority, medium priority and low priority according to the first expected connection probability threshold and the second expected connection probability threshold; setting redial rules for the to-be-called numbers of different priorities, the redial rules including the number of redials and the redial interval; calculating the human resource capacity and system load capacity of the call center, and determining the maximum number of outbound call tasks that can be executed in each time period; generating a dynamic outbound call strategy according to the maximum number of outbound call tasks, the priority of the to-be-called number, the number of redials and the redial interval; setting a dynamic priority adjustment mechanism to prepare a real-time adjustment strategy for the tasks in the outbound call task queue; the calculation formula for the maximum number of outbound call tasks is as follows:

[0011] M t =min(A t ×P a ,S t )×(1-B t )×E t

[0012] Among them, M t is the maximum number of outbound call tasks that can be executed within time period t, A t is the number of seats available in time period t, P a is the average productivity coefficient of each agent, S t is the maximum concurrent call processing capacity of the system in time period t, B t is the buffer coefficient reserved in time period t, E t is the efficiency adjustment coefficient for time period t.

[0013] As a preferred solution of the method for improving the outbound call connection rate of the call center described in the present invention, the generation of a dynamic outbound call strategy includes the following steps: sorting the to-be-called numbers according to their priorities, with high priorities being placed in front; sorting the to-be-called numbers within each priority level from high to low according to the expected connection probability; considering the optimal outbound call time window for each to-be-called number, allocating the to-be-called numbers to the corresponding time periods; for each time period, selecting a corresponding number of highly ranked numbers according to the maximum number of outbound call tasks; if the number of to-be-called numbers in a certain time period is less than the maximum number of outbound call tasks, supplementing them from the to-be-called numbers of the next priority; if the number of to-be-called numbers in a certain time period is greater than the maximum number of outbound call tasks, moving the excess to-be-called numbers to the next available time period; for numbers that need to be redialed, calculating the next redial time according to their priorities and redial rules, and inserting them into the corresponding time period; repeatedly executing the time period task allocation process until a preset termination condition is met, the preset termination condition includes that the number of outbound call tasks for all time periods reaches the maximum number of outbound call tasks, or all to-be-called numbers are arranged to the corresponding time period; converting the allocation result into an outbound call task queue.

[0014] As a preferred solution of the method for improving the outbound call connection rate of the call center described in the present invention, the method comprises the following steps: executing the outbound call task according to the outbound call task queue includes the following steps: initializing the outbound call task execution environment, obtaining the outbound call task of the current time period from the outbound call task queue; for each outbound call task, initiating a call within its optimal outbound call time window, starting a timer, and recording the call start time; real-time monitoring of the outbound call process, recording the call status, and performing corresponding processing according to the call status; adjusting the task priority in real time and updating the outbound call task queue according to the priority dynamic adjustment mechanism; real-time statistics of the outbound call volume, connection rate and average call duration in each time period, and dynamically adjusting the outbound call task execution speed according to the statistical results; for the outbound call tasks that have been executed, generating an outbound call task record and storing it in a database in real time; regularly checking the outbound call task queue, and marking the outbound call tasks that exceed the maximum number of redial times or the predetermined deadline as terminated; if the outbound call task is in a terminated state, removing it from the outbound call task queue and generating a terminated task analysis report.

[0015] As a preferred solution of the method for improving the outbound call connection rate of the call center described in the present invention, wherein: using the outbound call task records to update the outbound call connection prediction model and the outbound call strategy includes the following steps: regularly collecting and organizing the outbound call task records; merging the outbound call task records with the historical outbound call data to form an incremental training data set, and using the incremental training data set to update the outbound call connection prediction model; calculating the prediction accuracy of the updated outbound call connection prediction model within a preset time period, and comparing it with the outbound call connection prediction model currently in use; if the prediction accuracy of the updated outbound call connection prediction model is improved by more than a second preset threshold, then using the updated outbound call connection prediction model to replace the original outbound call connection prediction model; if the prediction accuracy of the updated outbound call connection prediction model is not improved by more than the second preset threshold, then The original outbound call connection prediction model is retained, and the reasons for the update failure are recorded; the optimal outbound call time window and expected connection probability of all numbers to be called are updated using the updated outbound call connection prediction model; the deviation between the actual connection rate of the outbound call task and the expected connection probability is analyzed, and the first expected connection probability threshold and the second expected connection probability threshold are adjusted according to the analysis results; the number of redials and the redial interval are optimized based on the actual outbound call results; the outbound call strategy is regenerated using the updated optimal outbound call time window, expected connection probability, and the optimized initial retry number allocation rule and retry interval rule; the newly generated outbound call strategy is applied to the next round of outbound call task queue generation process; the overall outbound call connection rate of the call center is regularly calculated and recorded, the optimization effect is evaluated, and an optimization report is formed.

[0016] In the second aspect, an embodiment of the present invention provides a call center outbound call connection rate improvement system, which includes a data collection module for collecting historical outbound call data of the call center; a model building module for analyzing historical outbound call data using a machine learning algorithm and establishing an outbound call connection prediction model; a calculation module for calculating the optimal outbound call time window and the expected connection probability for each number to be called based on the outbound call connection prediction model; a strategy formulation module for formulating an outbound call strategy according to the optimal outbound call time window and the expected connection probability, and generating an outbound call task queue according to the outbound call strategy; a task execution module for executing outbound call tasks according to the outbound call task queue and generating an outbound call task record; an update optimization module for updating the outbound call connection prediction model and the outbound call strategy using the outbound call task record, dynamically optimizing the outbound call process, and completing the improvement of the outbound call connection rate of the call center.

[0017] In a third aspect, an embodiment of the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program instructions are executed by the processor, the steps of the method for improving the outbound call connection rate of the call center as described in the first aspect of the present invention are implemented.

[0018] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program instructions are executed by a processor, the steps of the method for improving the outbound call connection rate of a call center as described in the first aspect of the present invention are implemented.

[0019] The beneficial effects of the present invention are as follows: the present invention analyzes historical outbound call data through a machine learning algorithm, establishes an outbound call connection prediction model, and improves the prediction accuracy of the outbound call connection rate. This data-driven method enables the call center to more accurately grasp the best outbound call timing and significantly improves the overall outbound call efficiency. The present invention introduces multidimensional feature analysis and explainability analysis to enhance the reliability and transparency of the prediction model. This not only helps call center managers understand the key factors that affect the outbound call connection rate, but also provides strong support for the continuous optimization of outbound call strategies. The present invention designs a dynamic outbound call strategy and priority adjustment mechanism, which can flexibly adjust task priorities and execution plans according to real-time outbound call results. This adaptive method ensures the efficient use of outbound call resources while improving the service quality for high-value customers. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 A method to improve the outbound call connection rate of the call center.

[0022] Figure 2 A method to improve the outbound call connection rate of the call center.

[0023] Figure 3 A method to improve the outbound call connection rate of the call center.

[0024] Figure 4 A diagram of computer equipment for a method of improving the outbound call completion rate of a call center. DETAILED DESCRIPTION

[0025] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0026] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0027] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0028] Example 1, reference Figure 1 to Figure 4 , which is the first embodiment of the present invention, and provides a method for improving the outbound call connection rate of a call center, comprising:

[0029] S1: Collect historical outbound call data of the call center.

[0030] It should be noted that the historical outbound call data includes but is not limited to outbound call time, call duration, connection status, user feedback information, user attributes, historical call records and environmental factors.

[0031] Furthermore, the historical outbound call data is divided into a training data set and a verification data set according to a preset ratio, and the present invention predicts that the ratio is 4:1.

[0032] S2: Use machine learning algorithms to analyze historical outbound call data and establish an outbound call connection prediction model.

[0033] Specifically, the following steps are included:

[0034] S2.1: Preprocess historical outbound call data.

[0035] Specifically, preprocessing includes data cleaning, missing value processing, and outlier identification.

[0036] S2.2: Extract multi-dimensional features from the pre-processed historical outbound call data.

[0037] It should be noted that multidimensional features include time features, user features and environmental features. Time features include the hours, day of the week, month and whether it is a weekday of the outbound call time. User features include the user's historical connection rate, average call duration, time interval since the last successful connection, and user attribute category. Environmental features include the weather type of the day, whether it is a statutory holiday, and local major event markers.

[0038] S2.3: Construct training samples, taking multi-dimensional features as input and the on-state as output label.

[0039] S2.4: Select multiple machine learning algorithms, and use the training samples to train each machine learning algorithm to obtain multiple candidate models.

[0040] It should be noted that machine learning algorithms include but are not limited to decision trees, random forests, and gradient boosting trees.

[0041] S2.5: Use the verification data set to calculate the connection probability prediction accuracy of each candidate model, and select the candidate model with the highest connection probability prediction accuracy as the outbound call connection prediction model.

[0042] S2.6: Perform explainability analysis on the outbound call connection prediction model to improve the transparency and explainability of the outbound call connection prediction model.

[0043] S3: Based on the outbound call connection prediction model, the optimal outbound call time window and the expected connection probability are calculated for each number to be called.

[0044] Specifically, the following steps are included:

[0045] S3.1: Collect multi-dimensional information of the number to be called.

[0046] It should be noted that the multi-dimensional information includes user historical call records, user attributes and current time.

[0047] S3.2: Extract features from the multi-dimensional information of the number to be called, and generate a feature vector that is consistent with the input format of the outbound call connection prediction model.

[0048] S3.3: Input the feature vector into the outbound call connection prediction model to obtain the expected connection probability at different time periods of the day.

[0049] S3.4: Calculate the average expected connection probability of all time periods of the day, and mark the time periods with higher expected connection probability as candidate outbound call time windows.

[0050] S3.5: Perform cluster analysis on the candidate outbound call time windows, identify and merge temporally continuous windows, and form an optimal outbound call time window.

[0051] Specifically, the candidate outbound call time window is converted into a time series in minutes, and each time point is expressed as the number of minutes from midnight of the day; a time interval threshold Δt is set to determine the continuity of the time window, and Δt is set to 30 minutes in the present invention; the time series is traversed, and adjacent time points with a time interval less than or equal to Δt are classified into the same cluster; for each cluster, its start time, end time and duration are calculated; a minimum window duration threshold Tmin and a maximum window duration threshold Tmax are set, and Tmin is set to 60 minutes and Tmax is set to 240 minutes in the present invention; and screen for durations greater than or equal to Tmin, as the effective outbound call time window; calculate the average expected connection probability in each effective outbound call time window; sort the effective outbound call time windows in descending order according to the average expected connection probability, and select the first window after sorting as the optimal outbound call time window; if the duration of the optimal outbound call time window exceeds the maximum window duration threshold Tmax, divide the optimal outbound call time window into N sub-windows of equal length, and the duration of each sub-window does not exceed Tmax; if there are multiple sub-windows, all of these sub-windows are regarded as the optimal outbound call time windows and are numbered in the original time order.

[0052] S3.6: Calculate the average expected connection probability within the optimal outbound call time window as the expected connection probability of the to-be-called number.

[0053] S3.7: The optimal outbound call time window, the expected connection probability and the to-be-called number are associated and stored for subsequent outbound call task scheduling.

[0054] S4: formulate an outbound call strategy based on the optimal outbound call time window and the expected connection probability, and generate an outbound call task queue based on the outbound call strategy.

[0055] Specifically, the following steps are included:

[0056] S4.1: Setting a first expected connection probability threshold and a second expected connection probability threshold to distinguish high priority numbers, medium priority numbers and low priority numbers.

[0057] Specifically, considering the business characteristics and historical data distribution of the call center, the first expected connection probability threshold and the second expected connection probability threshold are set by a dynamic adjustment method. The first expected connection probability threshold and the second expected connection probability threshold are initially set based on the statistical distribution of historical data, and then dynamically adjusted according to the actual outbound call results. After each preset number of outbound call tasks are completed, the deviation between the actual connection rate and the expected connection probability is calculated. If the deviation exceeds the preset range, the first expected connection probability threshold and the second expected connection probability threshold are adjusted accordingly to ensure the accuracy of priority classification and the effectiveness of the outbound call strategy.

[0058] S4.2: According to the first expected connection probability threshold and the second expected connection probability threshold, the to-be-called numbers are classified into high priority, medium priority and low priority.

[0059] Specifically, numbers whose expected connection probability is higher than or equal to the second expected connection probability threshold are marked as high priority; numbers whose expected connection probability is between the first expected connection probability threshold and the second expected connection probability threshold are marked as medium priority; numbers whose expected connection probability is lower than or equal to the first expected connection probability threshold are marked as low priority.

[0060] S4.3: Set redial rules for the waiting numbers of different priorities, including the number of redial times and the redial interval time.

[0061] For example, for high-priority numbers, the maximum number of redials is set to 6 times, the initial redial interval is 30 minutes, and the interval increases by 30 minutes after each redial failure; for medium-priority numbers, the maximum number of redials is set to 4 times, the initial redial interval is 60 minutes, and the interval increases by 60 minutes after each redial failure; for low-priority numbers, the maximum number of redials is set to 2 times, and the redial interval is 120 minutes.

[0062] S4.4: Calculate the human resource capacity and system load capacity of the call center and determine the maximum number of outbound call tasks that can be performed in each time period.

[0063] Preferably, the calculation formula for the maximum number of outbound call tasks is as follows:

[0064] M t =min(A t ×P a ,S t )×(1-B t )×E t

[0065] Among them, M t is the maximum number of outbound call tasks that can be executed within time period t, A t is the number of seats available in time period t, P a The average productivity coefficient of each agent (the average number of outbound call tasks that can be handled per hour), S t is the maximum concurrent call processing capacity of the system in time period t, B t is the buffer factor reserved in time period t (used to handle unexpected situations and answer inbound calls, with a value range of 0-1), t is the efficiency adjustment coefficient for time period t (derived from historical data analysis, used to adjust efficiency differences in different time periods).

[0066] S4.5: Generate a dynamic outbound call strategy based on the maximum number of outbound call tasks, the priority of the number to be called, the number of redial times, and the redial interval time.

[0067] Preferably, the generation process of the dynamic outbound call strategy includes the following steps: sorting the to-be-called numbers according to their priorities, with high priority numbers in front; sorting the to-be-called numbers within each priority level from high to low according to the expected connection probability; considering the optimal outbound call time window for each to-be-called number, allocating the to-be-called numbers to the corresponding time periods; for each time period, selecting a corresponding number of highly ranked numbers based on the maximum number of outbound call tasks; for numbers that need to be redialed, calculating the next redial time based on their priority and redial rules, and inserting them into the corresponding time period; if the number of to-be-called numbers in a certain time period is less than the maximum number of outbound call tasks, supplementing them from the to-be-called numbers of the next priority level; if the number of to-be-called numbers in a certain time period is greater than the maximum number of outbound call tasks, moving the excess to-be-called numbers to the next available time period; repeating the time period task allocation process until the preset termination condition is met; converting the allocation result into an outbound call task queue, each task containing the to-be-called number, priority, expected connection probability, and optimal outbound call time window.

[0068] The predetermined termination conditions include: the number of outbound call tasks in all time periods reaches the maximum number of outbound call tasks, or all to-be-called numbers are arranged in the corresponding time periods.

[0069] S4.6: Set up a dynamic priority adjustment mechanism to prepare real-time adjustment strategies for tasks in the outbound call task queue.

[0070] It should be noted that the dynamic adjustment mechanism of priority includes the following contents: for numbers that are not connected, for high-priority tasks, if the number of consecutive unconnected times reaches 60% of its maximum number of redial times, the priority will be reduced by one level; for medium-priority tasks, if the number of consecutive unconnected times reaches 80% of its maximum number of redial times, the priority will be reduced by one level; for low-priority tasks, if the number of consecutive unconnected times reaches its maximum number of redial times, it will be marked as pending evaluation; for numbers that are successfully connected, if the call duration exceeds the first preset threshold, the following operations are performed according to the current priority: low-priority tasks are promoted to medium priority, medium-priority tasks are promoted to high priority, and high-priority tasks remain unchanged, but are ranked higher in the same level.

[0071] S5: Execute outbound call tasks according to the outbound call task queue and record the results of each outbound call in real time.

[0072] Specifically, the following steps are included:

[0073] S5.1: Initialize the outbound call task execution environment and obtain the outbound call tasks in the current time period from the outbound call task queue.

[0074] S5.2: For each outbound call task, initiate a call within its optimal outbound call time window, start a timer, and record the call start time.

[0075] S5.3: Monitor the outbound call process in real time and record the call status.

[0076] It should be noted that the call status includes dialing, ringing, connected, not connected, busy, no one answered, wrong number, etc.

[0077] S5.4: Perform corresponding processing according to the call status.

[0078] For example, if the call status is connected, the connection time and call duration are recorded, user feedback information is collected, and user characteristics are updated; if the call status is connected and the user indicates that no contact is needed anymore, the relevant number is removed from the outbound call task queue and marked as completed; if the call status is connected and the user requests to reschedule the contact, the outbound call task is rescheduled according to the user-specified time, and its position in the outbound call task queue is adjusted accordingly; if the call status is not connected, the next redial time is set according to the redial rule set in S4.3, the task priority is updated according to the dynamic adjustment mechanism of S4.6, and the task is reinserted into the outbound call task queue; if the call status is busy, no one answers, or the number is wrong, the relevant task is marked as pending verification and temporarily removed from the active queue; for the pending verification status, the system automatically generates a list of pending verification tasks for manual review and processing.

[0079] S5.5: According to the dynamic priority adjustment mechanism, adjust the task priority in real time and update the outbound call task queue.

[0080] S5.6: Real-time statistics of outbound call volume, connection rate and average call duration in each time period, and dynamic adjustment of outbound call task execution speed based on statistical results.

[0081] S5.7: For the outbound call tasks that have been executed, an outbound call task record is generated and stored in the database in real time.

[0082] It should be noted that the outbound call task records include information such as outbound call time, call status, call duration, user feedback, and task priority changes.

[0083] S5.8: Regularly check the outbound call task queue and mark the outbound call tasks that exceed the maximum number of redial times or the predetermined deadline as terminated.

[0084] S5.9: If the outbound call task is in a terminated state, it is removed from the outbound call task queue and a terminated task analysis report is generated.

[0085] S6: Use the outbound call task records to update the outbound call connection prediction model and outbound call strategy, dynamically optimize the outbound call process, and improve the outbound call connection rate of the call center.

[0086] Preferably, outbound call task records are collected and sorted regularly; the outbound call task records are merged with the historical outbound call data of S1 to form an incremental training data set, and the incremental training data set is used to re-execute S2 to update the outbound call connection prediction model; the prediction accuracy of the updated outbound call connection prediction model within a preset time period is calculated, and compared with the outbound call connection prediction model currently in use; if the prediction accuracy of the updated outbound call connection prediction model is improved by more than a second preset threshold value (such as 1 / 4), the updated outbound call connection prediction model is used to replace the original outbound call connection prediction model; if the replacement standard is not met (the prediction accuracy improvement does not exceed the second preset threshold value), the original outbound call connection prediction model is retained, and the reason for the update failure is recorded; the updated outbound call connection prediction model is used to re-execute S3 to update the optimal outbound call time window and the expected connection probability of all numbers to be called; the actual connection rate of the outbound call task is analyzed against the expected Deviation of connection probability, adjust the first expected connection probability threshold and the second expected connection probability threshold in S4.1 according to the analysis results; optimize the number of redials and the redial interval in S4.3 based on the actual outbound call results; re-execute S4 using the updated optimal outbound call time window, expected connection probability, and the optimized initial retry count allocation rule and retry interval rule to generate a new outbound call strategy; apply the newly generated outbound call strategy to the next round of outbound call task queue generation process; regularly calculate and record the overall outbound call connection rate of the call center, evaluate the optimization effect, and form an optimization report; set an automated optimization trigger mechanism, including setting a fixed time interval trigger (such as weekly), a performance indicator trigger threshold (such as a 5% decrease in the connection rate), and a data volume trigger threshold (such as the number of new outbound call task records reaching 10% of the original data set). When any trigger condition is met, the optimized outbound call process of S6 is automatically executed.

[0087] Furthermore, the present embodiment also provides a call center outbound call connection rate improvement system, including a data collection module for collecting historical outbound call data of the call center; a model building module for analyzing historical outbound call data using a machine learning algorithm and establishing an outbound call connection prediction model; a calculation module for calculating the optimal outbound call time window and expected connection probability for each number to be called based on the outbound call connection prediction model; a strategy formulation module for formulating an outbound call strategy based on the optimal outbound call time window and the expected connection probability, and generating an outbound call task queue based on the outbound call strategy; a task execution module for executing outbound call tasks according to the outbound call task queue and generating an outbound call task record; an update optimization module for updating the outbound call connection prediction model and the outbound call strategy using the outbound call task record, dynamically optimizing the outbound call process, and completing the improvement of the call center's outbound call connection rate.

[0088] This embodiment also provides a computer device, which is suitable for the method of improving the outbound call connection rate of a call center, and includes a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the method of improving the outbound call connection rate of a call center as proposed in the above embodiment.

[0089] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0090] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the method for improving the outbound call connection rate of a call center as proposed in the above embodiment is implemented.

[0091] In summary, the present invention analyzes historical outbound call data through machine learning algorithms, establishes an outbound call connection prediction model, and improves the prediction accuracy of the outbound call connection rate. This data-driven method enables the call center to grasp the best outbound call timing more accurately, significantly improving the overall outbound call efficiency. The present invention introduces multidimensional feature analysis and explainability analysis to enhance the reliability and transparency of the prediction model. This not only helps call center managers understand the key factors affecting the outbound call connection rate, but also provides strong support for the continuous optimization of outbound call strategies. The present invention designs a dynamic outbound call strategy and priority adjustment mechanism, which can flexibly adjust the task priority and execution plan according to the real-time outbound call results. This adaptive method ensures the efficient use of outbound call resources while improving the service quality for high-value customers.

[0092] Example 2, reference Figure 1 to Figure 4 , which is the second embodiment of the present invention, and this embodiment provides a method for improving the outbound call connection rate of a call center. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation experiments.

[0093] To verify the effectiveness of the present invention, a three-month simulation experiment was conducted in a medium-sized call center with 200 seats, as follows:

[0094] First, the call center collected historical outbound call data from the past 12 months, totaling about 1.5 million records. These data include outbound call time, call duration, connection status, user feedback, etc. The data set was divided into a training set (1.2 million records) and a validation set (300,000 records) in a 4:1 ratio.

[0095] Next, the data was preprocessed, including cleaning outliers and processing missing values. After processing, the amount of valid data was 1.47 million, accounting for 98% of the original data. 28 features were extracted from these data, including 7 time features, 15 user features, and 6 environment features.

[0096] Subsequently, the model was trained using three algorithms: decision tree, random forest, and gradient boosting tree. After testing on the validation set, the gradient boosting tree model performed best, with a prediction accuracy of 83.2%. An interpretability analysis of the model revealed that the user's historical connection rate, outbound call time period, and user attributes were the three main factors affecting the connection probability.

[0097] Based on the trained model, the optimal outbound call time window and expected connection probability were calculated for the call center's current 500,000 waiting numbers. The results show that the optimal outbound call time for about 60% of the numbers is concentrated between 10:00-12:00 and 14:00-16:00 on weekdays. Numbers with an expected connection probability greater than 65% are marked as high priority, those between 45%-65% are marked as medium priority, and those below 45% are marked as low priority.

[0098] Specifically, different redial rules are set according to priority: high priority can redial up to 6 times, with an initial interval of 30 minutes; medium priority can redial up to 4 times, with an initial interval of 60 minutes; low priority can redial up to 2 times, with an interval of 120 minutes. Considering the manpower and system capacity of the call center, the maximum number of outbound call tasks per hour is set to 3,600.

[0099] Preferably, a dynamic priority adjustment mechanism is implemented. Experimental data shows that about 5% of low-priority tasks are promoted to medium priority due to successful connection and call duration exceeding 5 minutes, and 3% of medium-priority tasks are promoted to high priority. At the same time, about 8% of high-priority tasks are downgraded to medium priority due to multiple failures to connect.

[0100] During the experiment, the system executed about 2.7 million outbound calls, an average of 30,000 calls per day. Real-time records show that the overall connection rate increased from 65% before the experiment to 86%, an increase of 21%. The average connection rate for high-priority tasks reached 85%, medium-priority tasks reached 53%, and low-priority tasks reached 38%.

[0101] It is worth noting that the prediction model is updated every week with newly added outbound call task records. In the 8th week of the experiment, the prediction accuracy of the updated model increased to 85.7%, exceeding the preset 1% update threshold, so the new model was adopted.

[0102] In addition, it was observed that the average deviation between the actual connection rate and the expected connection probability was 4.3%. Based on this finding, the priority division threshold was adjusted in week 6, raising the high priority threshold to 68% and the low priority threshold to 48%. This adjustment made the task allocation more reasonable and further improved the overall efficiency.

[0103] Preferably, the comparison indicators of the present invention and the traditional method are shown in Table 1.

[0104] Table 1 Comparison indexes of the present invention and the traditional method

[0105] index The present invention Traditional methods Overall connection rate 86％ 65％ High priority connection rate 85％ 62.4％ Average call duration 4.2 minutes 3.5 minutes Seat Utilization 78％ 65％ Customer Satisfaction 83％ 75％ Prediction model accuracy 85.7％ not applicable Average outbound calls per hour 3400 times 2800 times Replay times (average) 3.5 times 2.8 times Task completion rate 93％ 82％ Operating cost reduction ratio 15％ Benchmarks

[0106] As can be seen from Table 1, the method of the present invention is significantly superior to the traditional method in all key indicators. The overall connection rate increased by 21%, and the connection rate of high-priority customers increased by 22.6%. The average call duration increased by 20%, indicating that the quality of customer communication has improved. The seat utilization rate increased by 20%, showing the improvement in resource allocation efficiency. Customer satisfaction increased by 8%, reflecting the overall improvement in service quality. The prediction model introduced by the present invention has a high accuracy of 85.7%, while the traditional method lacks such prediction capabilities. The average outbound call volume per hour increased by 21.4%, while the average number of redials decreased by 20%, reflecting the optimization of the outbound call strategy. The average daily task completion rate increased by 13.4%, highlighting the improvement in work efficiency. Finally, the operating cost decreased by 15%, demonstrating the significant advantages of the present invention in improving efficiency while reducing costs.

[0107] In summary, by implementing the technical solution of the present invention, the outbound call connection rate of the call center is significantly improved and the resource utilization efficiency is greatly improved.

[0108] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for improving the outbound call connection rate of a call center, characterized by: include, Collecting historical outbound call data of the call center, including outbound call time, call duration, connection status, and user feedback information; Analyze the historical outbound call data using a machine learning algorithm to establish an outbound call connection prediction model; Based on the outbound call connection prediction model, the optimal outbound call time window and the expected connection probability are calculated for each number to be called; Formulate an outbound call strategy according to the optimal outbound call time window and the expected connection probability, and generate an outbound call task queue according to the outbound call strategy; Execute the outbound call task according to the outbound call task queue and generate an outbound call task record; The outbound call task record is used to update the outbound call connection prediction model and outbound call strategy, dynamically optimize the outbound call process, and improve the outbound call connection rate of the call center.

2. The method for improving the outbound call connection rate of a call center according to claim 1, characterized in that: The establishment of the outbound call connection prediction model comprises the following steps: Preprocessing the historical outbound call data; Extracting multidimensional features from the preprocessed historical outbound call data, wherein the multidimensional features include time features, user features, and environment features; Construct training samples, taking multi-dimensional features as input and the connection state as output label; Select multiple machine learning algorithms, and use training samples to train each machine learning algorithm to obtain multiple candidate models; The verification data set is used to calculate the connection probability prediction accuracy of each candidate model, and the candidate model with the highest connection probability prediction accuracy is selected as the outbound call connection prediction model; Conduct interpretable analysis on the outbound call connection prediction model.

3. The method for improving the outbound call connection rate of a call center according to claim 1, characterized in that: Calculating the optimal outbound call time window and the expected connection probability for each to-be-called number comprises the following steps: Collect multi-dimensional information of the number to be called, the multi-dimensional information including user historical call records, user attributes and current time; Extract features from the multi-dimensional information of the call number to generate a feature vector consistent with the input format of the outbound call connection prediction model; Input the feature vector into the outbound call connection prediction model to obtain the expected connection probability at different time periods of the day; Calculate the average expected connection probability for all time periods of the day, and mark the time periods with higher expected connection probability as candidate outbound call time windows; Perform cluster analysis on candidate outbound call time windows, identify and merge temporally continuous windows, and form the best outbound call time window; Set the maximum window duration threshold Tmax. If the duration of the optimal outbound call time window exceeds Tmax, the optimal outbound call time window is divided into N sub-windows of equal length, and the duration of each sub-window is less than or equal to Tmax. If there are multiple sub-windows, all of them are considered as the best outbound call time windows and numbered in the original time order; Calculate the average expected connection probability within the optimal outbound call time window as the expected connection probability of the to-be-called number; The optimal outbound call time window, the expected connection probability and the to-be-called number are associated and stored.

4. The method for improving the outbound call connection rate of a call center according to claim 1, characterized in that: The formulation of the outbound call strategy comprises the following steps: Setting a first expected connection probability threshold and a second expected connection probability threshold; According to the first expected connection probability threshold and the second expected connection probability threshold, the to-be-called numbers are classified into high priority, medium priority and low priority; Setting redial rules for to-be-called numbers of different priorities, the redial rules including the number of redial times and the redial interval time; Calculate the call center's human resource capacity and system load capacity to determine the maximum number of outbound call tasks that can be performed in each time period; Generate dynamic outbound call strategies based on the maximum number of outbound call tasks, priority of the number to be called, number of redials, and redial interval; Set up a dynamic priority adjustment mechanism to prepare real-time adjustment strategies for tasks in the outbound call task queue; The calculation formula for the maximum number of outbound call tasks is as follows: M t =min(A t ×P a ,S t )×(1-B t )×E t Among them, M t is the maximum number of outbound call tasks that can be executed within time period t, A t is the number of seats available in time period t, P a is the average productivity coefficient of each agent, S t is the maximum concurrent call processing capacity of the system in time period t, B t is the buffer factor reserved in time period t, E t is the efficiency adjustment coefficient for time period t.

5. The method for improving the outbound call connection rate of a call center according to claim 4, characterized in that: Generating a dynamic outbound call strategy comprises the following steps: Sort by priority of the numbers to be called, with higher priority numbers at the front; The waiting-to-call numbers within each priority level are sorted from high to low according to the expected connection probability; Consider the optimal outbound call time window for each waiting number and assign the waiting number to the corresponding time period; For each time period, select a corresponding number of high-ranking numbers based on the maximum number of outbound call tasks; If the number of waiting numbers in a certain time period is less than the maximum number of outbound call tasks, the waiting numbers of the next priority level will be used to supplement the number; If the number of waiting numbers in a certain time period is greater than the maximum number of outbound call tasks, the excess waiting numbers will be moved to the next available time period; For numbers that need to be redialed, the next redial time is calculated based on its priority and redial rules, and inserted into the corresponding time period; Repeat the time period task allocation process until a preset termination condition is met, wherein the preset termination condition includes that the number of outbound call tasks in all time periods reaches the maximum number of outbound call tasks, or all to-be-called numbers are arranged in the corresponding time period; Convert the allocation results into an outbound call task queue.

6. The method for improving the outbound call connection rate of a call center according to claim 1, characterized in that: Executing the outbound call task according to the outbound call task queue comprises the following steps: Initialize the outbound call task execution environment and obtain the outbound call tasks in the current time period from the outbound call task queue; For each outbound call task, initiate a call within its optimal outbound call time window, start a timer, and record the call start time; Monitor the outbound call process in real time, record the call status, and perform corresponding processing according to the call status; According to the dynamic priority adjustment mechanism, the task priority is adjusted in real time and the outbound call task queue is updated; Real-time statistics of outbound call volume, connection rate and average call duration in each time period, and dynamic adjustment of outbound call task execution speed based on statistical results; For completed outbound call tasks, an outbound call task record is generated and stored in the database in real time; Regularly check the outbound call task queue and mark the outbound call tasks that exceed the maximum number of redial times or the scheduled deadline as terminated; If the outbound call task is in a terminated state, it will be removed from the outbound call task queue and a terminated task analysis report will be generated.

7. The method for improving the outbound call connection rate of a call center according to claim 6, characterized in that: The updating of the outbound call connection prediction model and the outbound call strategy by using the outbound call task record comprises the following steps: Regularly collect and organize outbound call task records; Merge the outbound call task records with the historical outbound call data to form an incremental training data set, and use the incremental training data set to update the outbound call connection prediction model; Calculate the prediction accuracy of the updated outbound call connection prediction model within a preset time period and compare it with the currently used outbound call connection prediction model; If the prediction accuracy of the updated outbound call connection prediction model is improved by more than a second preset threshold, the updated outbound call connection prediction model is used to replace the original outbound call connection prediction model; If the prediction accuracy of the updated outbound call connection prediction model does not increase by more than the second preset threshold, the original outbound call connection prediction model is retained and the reason for the update failure is recorded; Using the updated outbound call connection prediction model, the optimal outbound call time window and expected connection probability of all waiting numbers are updated; Analyze the deviation between the actual connection rate of the outbound call task and the expected connection probability, and adjust the first expected connection probability threshold and the second expected connection probability threshold according to the analysis result; Optimize the number of redials and the redial interval based on actual outbound call results; Regenerate the outbound call strategy using the updated optimal outbound call time window, expected connection probability, and optimized initial retry count allocation rule and retry interval rule; Apply the newly generated outbound call strategy to the next round of outbound call task queue generation process; Regularly calculate and record the overall outbound call connection rate of the call center, evaluate the optimization effect, and form an optimization report.

8. A call center outbound call connection rate improvement system, based on the call center outbound call connection rate improvement method according to any one of claims 1 to 7, characterized in that: Also includes, A data collection module is used to collect historical outbound call data of the call center, wherein the historical outbound call data includes outbound call time, call duration, connection status, and user feedback information; A model building module, used to analyze the historical outbound call data using a machine learning algorithm to establish an outbound call connection prediction model; A calculation module, used for calculating the optimal outbound call time window and the expected connection probability for each number to be called based on the outbound call connection prediction model; A strategy formulation module, used to formulate an outbound call strategy according to the optimal outbound call time window and the expected connection probability, and generate an outbound call task queue according to the outbound call strategy; A task execution module, used for executing the outbound call task according to the outbound call task queue and generating an outbound call task record; The update optimization module is used to update the outbound call connection prediction model and outbound call strategy using the outbound call task record, dynamically optimize the outbound call process, and improve the outbound call connection rate of the call center.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method for improving the outbound call connection rate of a call center according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for improving the outbound call connection rate of a call center according to any one of claims 1 to 7 are implemented.

Images